The next problem was to give a reasonable account of the extraordinary observation. Both the Italians and the French were not long in arriving at a correct explanation, which was based upon the facts observed when two vertical tubes are joined at the bottom and filled with different liquids. A heavy liquid in one tube can counterbalance a longer column of a lighter liquid in the other tube. If two tubes could be constructed each one hundred miles high, they would reach nearly to the top of the atmosphere. Now, if one of the tubes was left full of air, and if instead of air mercury could be poured into the other tube, a very short column of mercury would balance the whole column of air in the other tube. The length of the mercury column would be as many times shorter than the air column as the mercury is heavier than air. The two tubes would in fact form and any change in the amount of air in the air tube would be indicated immediately by a rise or fall of the balancing column of mercury. [Illustration: Fig. 34.--The pressure of the air upon the mercury in the short open tube keeps up a column of mercury about 30 inches long in the closed tube.] In a barometer it is therefore sufficient to close the tube containing the mercury at a height of about three feet, so that a tube one hundred miles high is not required there. Nor is such a long tube required on the other side either. For, as we have already seen, the pressure of a liquid simply depends upon the height of its surface above the bottom upon which it presses, and as the width of the tube makes no difference the air tube may be made as wide as we please, or may be taken away altogether, so long as we prevent the air bubbling up into the vacuum. This can be done by bending round the lower end of the mercury column until it points upwards. The mercury then rises and falls both in the long tube and in the short tube, and the pressure of the air upon the short tube is measured by the difference of level of the mercury in the two tubes. It would, of course, be more convenient if the height of the mercury by itself indicated the pressure of the air, and this can be secured by dipping the lower end of the mercury tube into a large vessel full of mercury, so that its level does not perceptibly change when a little mercury flows in or out of the barometer tube. This is, in fact, the modern form of the mercury barometer. Torricelli himself was quite aware of the importance of his experiment. He wrote, “I do not mean simply to produce a vacuum, but to make an instrument which shows the changes of the air, now heavy and dense, and now lighter and thin.”